May, 1960 ExPERIMENTAL ProrriNG 143 
Reference 5 
  
  
  
Fic. 3. The effect of the shifting of the 
reference plane on systematic errors. 
  
  
  
twelve plots. On an average the positive elevation errors in these twelve plots exceed 
the negative elevation errors by 1.54 m. The negligible (—0.29 m) average value of 
the prevailing negative errors in the single plot is purely accidental, as will be shown 
later. 
Failure to shift the reference plane properly is the principal cause of the large 
elevation errors in the results submitted by many participants. The results are listed 
in Table I. 
KNowN SvsrEMATIC Ennons 
For spot elevations the known systematic errors that should be corrected in current 
photogrammetric operations are errors caused by lens distortion and errors caused by 
earth curvature. Other systematic errors that could be considered are those caused 
by atmospheric refraction and those by imperfections in the plotters. So far the only 
errors that are usually corrected automatically during the plotting operation or during 
the production of diapositives are those caused by lens distortion. Because of this, 
earth curvature was almost forgotten and several participants assigned the discrepancy 
in the central control point to the wrong ground elevation or to misreading, and not 
to the natural effect of earth curvature. From the submitted spot elevations it was in 
the main quite evident whether or not elevations had been corrected for the effect of 
earth curvature. Consequently, the required corrections were applied where needed, 
together with the correction for atmospheric refraction. Thus, data were brought 
to a common basis for comparison and they agreed quite well with each other within 
each instrument group, as shown in Table IL 
Apart from the correction for systematic errors, the reference plane was shifted 
to equalize the maximum positive and negative discrepancies, as explained previously, 
by considering only five control points.